Analytics and Tracking

Anonymized Query

Also called Anonymised query, hidden query, filtered query

A search made by too few people for Google to list it, withheld so the searcher cannot be identified.

Quick facts: Anonymized Query

Category
Analytics and Tracking
Also called
Anonymised query, hidden query, filtered query
Level
Intermediate
Affects
Reported query totals, keyword research, long-tail visibility
Where to see it
Google Search Console (Performance), Search Console API, BigQuery export
In this article4
  1. How an anonymized query works
  2. Why anonymized queries matter
  3. Where anonymized queries go wrong
  4. How to act on it

How an anonymized query works

Search Console reports the searches that brought impressions and clicks to your site, but it does not report all of them. When a search has been made by very few people, Google withholds that row completely. The reasoning is straightforward: an unusual phrase attached to a small site could be traced back to the person who typed it, especially when the phrase contains a name, a location, a phone number or a medical worry.

The withheld impressions and clicks are not deleted. They remain inside the totals at the top of the report and inside the page, country and device breakdowns. Only the query dimension loses them. That is why the numbers stop matching the moment you switch to the queries tab: the totals include everything, while the rows include only what can be shown without exposing an individual.

Why anonymized queries matter

Mainly because of what people conclude when they have not heard of them. Someone comparing a query export against the total in the same report assumes the data is broken, or that whoever supplied it has quietly filtered something out. Nothing is broken and nothing has been removed.

They also put a hard limit on what Search Console can do for keyword research. The long tail — the wordy, occasional, oddly phrased searches that make up a large share of a small site’s demand — is exactly the part most likely to be withheld. A new site, a site in a small market, or a site picking up Nepali-language searches with modest volume loses a bigger share of its query detail than a large English-language site does.

Where anonymized queries go wrong

The usual mistake is treating the query list as a full inventory and building a content plan from it alone. A phrase missing from the list is not evidence that nobody searched it; it may simply be evidence that few people did.

The second mistake is arithmetic. Adding up query rows and presenting the result as your search traffic understates every period, and because the withheld share changes from month to month, the trend line you draw from those sums is unreliable as well. Report the totals row and let the query rows explain it.

How to act on it

Accept the gap and design your reporting around it. Use the page dimension whenever a figure has to be complete, because pages are not withheld, and use the query dimension for direction rather than for accounting. When you need the shape of the long tail, work in groups: filter queries containing a word, read how many rows come back and how they move over time, rather than studying single rows.

For real demand research, pair this with the search terms your ads collect and with a dedicated tool, then confirm the picture against the Search Console performance report. Together those sources cover most of what any one of them hides. If the whole point of the exercise is a content plan, treat the exported queries as a starting shortlist for proper keyword research, never as the finished list.

Do and do not

Do

  • Report the totals row, not the sum of queries
  • Use page data when a figure must be complete
  • Group queries by word to read the long tail

Do not

  • Read a missing query as proof nobody searched it
  • Assume the API returns the withheld rows
  • Build a content plan from exported queries alone

Questions people ask about this

Why do my Search Console query rows not add up to the total?

Because rare searches are withheld from the query list to protect the people who made them, while still being counted in the site total. The gap is normal and it changes size from period to period. Smaller sites and newer sites usually see a larger gap, since more of their searches fall below the threshold for reporting.

Can I get the hidden searches from the Search Console API instead?

No. The withholding happens before the data is served, so the API, Looker Studio and the BigQuery export all return the same filtered query rows as the interface. What the API gives you is more rows overall and easier automation, not access to the searches Google has chosen not to disclose.

Does an anonymized query mean I am losing rankings?

No. It only means the search was uncommon. Your page can rank in first place for a phrase and still never see that phrase listed, simply because too few people typed it. Judge visibility from page-level impressions and clicks, which include every search, rather than from the presence or absence of a row.

Related terms

Found this useful?

Share it, or ask an AI to summarise it

Back to the glossary

Knowing the term is the easy part

Applying it to your own site and budget is the work. Book a call and I will tell you what actually applies to you.